The present disclosure relates to Application Performance Management/Monitoring (APM), and more particularly to the application of APM in abnormality detection.
Typical application abnormality detection solutions are threshold based. Threshold settings can be difficult to configure. Threshold based application abnormality detection can be resource based or experience based. Resource based thresholds can include target CPU usage, target memory utilization, target threshold pool utilization, etc. Experience based thresholds can include target response time, target throughput, target packet drop rate, etc. Moreover, expert knowledge is needed to determine a threshold for dynamic workloads.